PROJECT SUMMARY/ABSTRACT Lesbian, gay, bisexual, and transgender (LGBT) adolescents and adults, when compared to similar others, are at increased risk for suicide attempts and exposure to violence victimization. Suicide mortality is also elevated, especially among women, though research studies are sparse. Accordingly, NIH has formally designated sexual/gender minorities a health disparity population. Yet there remain large gaps in the scientific knowledge base that undermine development, dissemination and implementation of effective public health interventions to reduce risk in this vulnerable population. These include whether rates of homicide are similarly elevated among LGBT persons, whether suicide rates are elevated among the transgender population, what role gender and race/ethnicity play in shaping risk, as well as clarification of the contextual factors surrounding these deaths, including missed opportunities for intervention. The primary goal of the project is to identify patterns of risk and to characterize proximal factors associated with violent death due to suicide and homicide among LGBT individuals. Drawing from Minority Stress Theory and, for suicides, the `ideation-to-action' framework, we hypothesize that anti-gay stigma and discrimination creates a vulnerability for violent death among LGBT adolescents and adults. Capitalizing on new information classifying deaths for LGBT status in recent years of the National Violent Death Reporting System (NVDRS), we will make novel use of supervised machine learning techniques to recapture LGBT status among deaths in those 12 years and older in the 2003-2015 NVDRS (> 200,000 suicides, homicides, or homicide-suicides). Similar techniques will also be used to code NVDRS death narratives for co-occurring stigma-based contextual factors (e.g., gay disclosure, familial rejection), risk profiles (e.g. mental health, substance use), and recent services use in the immediate circumstances surrounding the death. This will allow us to achieve two study aims. For Aim 1, we will use NVDRS data and exogenous information estimating LGBT population size to generate state-level estimates of suicide and homicide rates, adjusted for population composition, among individuals 12 years and older. We hypothesize that LGBT population risk estimates for suicide and homicide will exceed those of non-LGBT populations. For Aim 2, we will identify sexual orientation and gender minority-linked differences in predisposing and proximal factors associated with suicide and homicide, including indicators of mental health and substance use, patterns of recent health care access, and characteristics of death circumstances, including HIV infection status. Consistent with theoretical predictions, we hypothesize that recent discrimination experiences will be strongly associated with sexual orientation and transgender status, that homicide deaths will more likely show evidence of hate-crime-related contextual factors, and that HIV-related issues will be an important though downward trending precipitant of suicide among sexual minority men. We also anticipate gender and ethnic/racial differences in these effects that will have important implications for future risk reducing interventions. Information obtained will greatly aid achieving the American Foundation for Suicide Prevention Project 2025 goal of reducing the suicide rate 20% by 2025 and NIMH's Zero Suicide efforts by providing interventionists with data to design screening and intervention efforts for the LGBT population.